How to Make UGC Ads in Multiple Languages (2026 Playbook)
Subtitles vs dubbing vs native-language creative, the performance lift each one earns, and the real per-market cost.
HTHighQualityUGC Team10 min read
Subtitles vs dubbing vs native-language creative, the performance lift each one earns, and the real per-market cost.
Your winning ad works in one language. Every other market is getting a subtitled version of it and converting worse.
That gap is fixable now for less than the price of dubbing, and it does not require finding a creator in each country. Below is what the numbers say about native-language creative, why dubbed UGC breaks in a way dubbed brand film does not, what actually changes per market beyond the words, and the per-market cost math.
Most brands expanding to a second market do the same thing. They take the ad that worked, add translated captions, and point it at Germany or Mexico or France.
Then CPMs look fine, CTR is soft, and the market gets written off as "not a fit for us". It usually was a fit. The ad just sounded foreign in a format whose entire premise is that a normal local person is talking to you.
Only 25.9% of internet users speak English, and 61% of everyone who localizes video stops at subtitles while 12% go as far as dubbing. Most of your category is doing the cheap version, which is exactly why the market is winnable.
Do not rebuild the whole ad to find out whether a market responds to native-language creative. Rebuild three seconds of it.
HighQualityUGC Team
Editorial
We run UGC ad tests daily and publish what holds up: real credit costs, real hook rates, no vendor fluff.
Platform-by-platform runtime benchmarks, why the two most-quoted numbers disagree, and what each extra second costs to render.
TikTok, Meta, YouTube, and EU AI Act labeling rules for synthetic ad creative, plus what stays exempt and how to comply.
The six production tells that make generated ads read as synthetic, plus the prompt and reference fixes for each one.
Take your best performer, rewrite only the hook in the target language, and produce that opening as native-language footage. Keep the rest of the ad exactly as it is, subtitles and all. Run it against the fully subtitled original in the same ad set for 7 days.
Watch hook rate first, not CTR. Hook rate is 3-second views divided by impressions, it moves days before CTR does, and it isolates the only variable you changed. If hook rate lifts, the market is a language problem and worth full localization. If it does not move, the market is an offer or targeting problem and translating the rest would have taught you nothing.
There are exactly three ways to ship a UGC ad in a second language, and they differ by which layer you replace: the text, the audio, or the whole performance.
| Approach | What gets replaced | Cost per 15s ad per market | Turnaround | Where it breaks |
|---|---|---|---|---|
| Translated subtitles | On-screen text only | $0 to $5 | Minutes | Audio is still foreign and the mouth still speaks English, so sound-on viewers hear the mismatch |
| AI or human dubbing | Audio track | $4 to $13 at $15 to $50 per finished minute | Hours to days | Lip movement no longer matches the words, which reads as fake in close-up UGC framing |
| Native-language creative | Script, voice and performance | About $2.85 rendered, or $100 to $500 per creator video | Minutes rendered, days with a creator | Needs a real script per market, not a translated one |
| Bilingual or mixed-language | Parts of voiceover and captions | Same as native | Same as native | Only works where the audience genuinely code-switches |
The pricing row is the part people get wrong. Professional dubbing runs $15 to $50 per finished minute, so dubbing a 15-second ad costs roughly $4 to $13 per language. Rendering a fresh native-language version costs about $2.85. The compromise option is the expensive one.
Subtitles are not worthless, they are just small. Adding them lifts views around 7.32% and average view time about 12%, which is a real gain on a creative you were running anyway. It is not a market-entry strategy.
The lift is not a rounding error, and it shows up in three separate places in the funnel at once.
The same tests recorded 24% lower CPC and 21% higher conversion rate. Cheaper clicks and better conversion from the same media budget is what makes localization a margin decision rather than a brand decision.
Buyer behavior explains why. 76% of online shoppers want product information in their own language, 65% prefer content in their language outright, and 72.1% of consumers spend most or all of their time on sites in their own language.
This is a hard floor, not a preference. Two in five people in that market are unreachable by your English creative no matter how good the hook is.
Audio specifically carries more weight than most media buyers assume. In TikTok research with NRG, Spanish-dominant speakers were 57% more likely to watch more ads from a brand after hearing Spanish voiceovers. The same work found English-only audiences responded just as well to bilingual Spanish and English ads, so mixed-language creative does not cost you the broad audience.
Localized video overall shows about a 40% increase in engagement against non-localized versions, and dubbed content averages a 45% views increase. The direction is consistent across every dataset. The size depends on how far you go.
Dubbing a cinematic brand film is fine. Viewers have accepted dubbed film for decades, and the framing is usually wide enough that mouths are not the story.
UGC is the opposite. It is shot at arm's length, the face fills the frame, and the entire persuasion mechanism is that this looks like a real person talking to their phone. The moment lip movement stops matching the audio, that mechanism is gone.
Viewers register mismatched lip movement within seconds even when they cannot say what is wrong, and retention curves for mismatched video consistently underperform synchronized versions. They do not report "bad dub". They report that it feels off, and they scroll.
Fixing lip sync alone does not rescue it either. The mouth is one signal among several, and frozen expression, flat eye behavior and emotional mismatch between voice and face all still read as synthetic. This is the same failure mode covered in why AI UGC ads look fake, arriving through a different door.
The practical conclusion: if the ad is a talking head in close framing, generate the performance in the target language instead of dubbing over a performance recorded in another one. If the ad is voiceover over product footage with no visible speaker, dubbing is fine, because there is no mouth to contradict.
Feeding your script to a translator and rendering the output produces a technically correct ad that sounds like nobody. Direct word-for-word translation fails on tone, idiom and humor, and native-written copy for each market consistently outperforms translated copy, especially on high-intent audiences.
Six things change per market. Only the first is language.
Do not localize alphabetically or by market size. Localize where you already have demand you are underserving, because that is where the lift shows up fastest and the answer arrives in days rather than a quarter.
1. Read your existing analytics first
Sort sessions by country and language. Any non-English market already sending you traffic with a conversion rate below your home market is a language problem you are paying for right now. That is market one.
2. Check CPM arbitrage in the ads manager
Pull CPM by country for your existing campaigns. Markets where CPMs run well below your home market let the same creative budget buy more impressions, so a smaller lift still pays back.
3. Confirm you can actually fulfil there
Shipping, returns, currency, support language. A localized ad pointing at a checkout that cannot take the order converts the click and loses the customer. Fix the path before you buy the traffic.
4. Start with one market and two angles
One language, two distinct angles, native scripts for both. Two variants is enough to separate a language effect from a creative effect, and small enough to kill in a week.
5. Judge against the local baseline, not your home numbers
A 1.2% CTR might be strong in one market and weak in another. Compare each market to its own first-week baseline, never to your best-performing home creative.
6. Only then add the next language
Markets compound badly when launched together. Five at once means five sets of results you cannot attribute and no clean read on whether localization or the offer moved the number.
The old blocker was sourcing. Finding, briefing, and paying a native creator per market at $100 to $500 per video is why most brands never got past subtitles, and the full breakdown of what a UGC video costs shows how fast that scales when you multiply it by markets.
Generated actors remove the sourcing step, which changes the sequencing. Here is the build order that keeps quality per market instead of producing five weak ads at once.
1. Lock the concept in your strongest market first
Localize a proven winner, never an untested idea. If the ad has not earned its place in one market, translating it just distributes the loss.
2. Write the script natively per market, not translated
Same beats, same product moment, same length. New words, written the way a person in that market would actually say them. This is the step that produces the +38% CTR, and it is the step most workflows skip.
3. Pick an actor who fits the market
Look and delivery should read as local. A generated actor lets you match the market without recasting, and keeping that actor identical across every ad in the market is what builds recognition. The method is in our guide to keeping AI actors consistent across videos.
4. Generate the performance in the target language, do not dub
The spoken line drives both the audio and the mouth, so the lip movement matches by construction. That is the difference between a localized ad and a dubbed one.
5. Review the storyboard before rendering anything
Wardrobe, setting, and on-screen text all carry cultural signal. Catching a wrong room or an overflowing German caption at the storyboard stage costs 15 credits. Catching it after the render costs the whole clip.
6. Have one native speaker read the script out loud
Not a translation check. A does-a-person-say-this check. It takes five minutes per market and catches the register problems no tool flags.
Step three is the one that quietly decides whether a market ever builds recognition, and the mechanics of holding a face steady across a whole batch are in our guide to keeping AI actors consistent across videos.
Using the published credit table where 1 credit is about $0.01, a 15-second ad on Seedance 2.0 is 1 storyboard (15 credits) plus 15 seconds of render at 18 credits per second (270 credits). That is 285 credits, about $2.85. A 30-second ad is 2 storyboards plus 540 credits of render, so 570 credits, about $5.70.
| Markets | Credits | Rendered cost | With creators at $100 to $500 per video | Dubbed at $15 to $50 per minute |
|---|---|---|---|---|
| 1 (home market only) | 285 | about $2.85 | $100 to $500 | not applicable |
| 3 languages | 855 | about $8.55 | $300 to $1,500 | $8 to $26 on top of the original |
| 5 languages | 1,425 | about $14.25 | $500 to $2,500 | $16 to $52 on top of the original |
| 5 languages, 2 angles each | 2,850 | about $28.50 | $1,000 to $5,000 | $32 to $104 on top of the original |
Read the last two columns together. Ten native-language creatives across five markets cost about $28.50 to render, which is less than dubbing the same ten clips and roughly 3% of what the creator route costs at the high end. You can check the per-model credit table on the pricing page if you want to run this on Kling or Veo instead, since render rates differ per model.
The cost argument is not the interesting part though. The interesting part is that at these numbers you can afford two angles per market instead of one, which is the difference between testing a market and guessing at it.
Native-language creative outperforms subtitled English by roughly 38% on CTR, 24% on CPC and 21% on conversion in non-English markets, and 40% of buyers in those markets were never reachable by the English version at all.
Dubbing is the trap in the middle. It costs more than rendering a native version, and in close-framed UGC it destroys the exact realism the format depends on.
At about $2.85 per 15-second creative, the question stopped being whether localization is worth the production budget. It is now whether you have written a real script for each market, which is the only part of this that still takes human judgment.
Three approaches exist. Translated subtitles cost almost nothing but leave the audio and lip movement foreign. Dubbing replaces the audio track at $15 to $50 per finished minute, roughly $4 to $13 for a 15-second ad, but breaks lip sync. Native-language creative rewrites the script per market and generates the performance in that language, costing about $2.85 per 15-second render with AI actors or $100 to $500 with a creator. Native-language creative is the only one of the three that lifts CTR, CPC and conversion rate together.
Neither, if the ad is a close-framed talking head. Subtitles lift views around 7.32% and view time around 12%, which is a small gain on creative you were already running. Dubbing breaks the lip sync that makes UGC read as real footage, and viewers register the mismatch within seconds even when they cannot articulate it. Dubbing is only safe when there is no visible speaker, such as voiceover over product footage. Otherwise generate the performance in the target language.
Localization tests in 2026 recorded 38% higher CTR, 24% lower CPC and 21% higher conversion rate for native-language creative versus the English equivalent in non-English markets. Localized video shows about 40% higher engagement than non-localized versions. On the buyer side, 76% of online shoppers want product information in their own language and 40% will never buy from a site that is not in their language.
No. Sourcing a native creator per market at $100 to $500 per video is why most brands stop at subtitles. Generated actors remove the sourcing step, so five markets cost about $14.25 in render credits instead of $500 to $2,500. What you still need per market is a natively written script, because word-for-word translation loses the tone, idiom and CTA phrasing that produce the performance lift.
Six things. The hook needs rewriting natively rather than translating. Text length shifts, with German running about 35% longer than English and French about 20% longer, which overflows on-screen captions. The CTA needs a local action phrase, since 'shop now' translates literally and lands flat. Pricing needs local currency and VAT-inclusive display in the EU. Social proof works better with local numbers and cities. And some markets, Germany in particular, expect subtitles as a default even on native-language audio.
Last updated August 1, 2026